Subject: Re: 1.4.2 Observations
To: Manuel Bouyer <bouyer@antioche.lip6.fr>
From: Thor Lancelot Simon <tls@rek.tjls.com>
List: tech-kern
Date: 03/30/2000 11:32:55
On Thu, Mar 30, 2000 at 03:32:22PM +0200, Manuel Bouyer wrote:
> On Wed, Mar 29, 2000 at 09:20:06PM +0200, Manuel Bouyer wrote:
> > My home machine, cyrix 133 with a promise Ultra/33 IDE adapter can handle
> > 4000 IDE IRQ/s without troubles. I'll have to run my bench on my system at
> > work tomorow.
> 
> Hum, I can't reproduce the problem using the raw device.
> This seems related to some usage pattern of the filesystem ...

I just discovered that myself.  Now it starts to make a lot more sense to
me that using LFS magically makes the problem go away.

Have we both arrived at the hypothesis that, sometime between the 
introduction of the pciide code in mid-1.3, and 1.4, the filesystem code
was broken so as to cause this problem on some systems (long hangs under 
heavy I/O load)?

Again I'd suspect disksort but LFS seems to feed the strategy routines just
like FFS does so...

-- 
Thor Lancelot Simon	                                      tls@rek.tjls.com
	"And where do all these highways go, now that we are free?"